With the ever-increasing wealth of data and computing power now available to seismologists, statistical data-driven techniques are becoming increasingly critical to exploit the information contained in such extensive datasets. Machine learning-based methods in particular are achieving notable results, and are now being applied across the entire spectrum of seismological tasks, including: seismic picking, event detection, event location, seismic imaging, and source modelling. Whilst early approaches often applied standard machine learning techniques, recently, increasingly targeted machine learning methods for seismology are being published, e.g., incorporating physical knowledge into the models.

This session aims to bring together seismologists who are developing machine learning techniques to improve seismological modelling or data recovery. We encourage contributions that facilitate discussion around how novel machine learning methods can advance the field of seismology. Submissions are not limited to classical seismological datasets but can also include more general applications (i.e., GNSS and InSAR data).